A3M Router – Parallel LLM routing across 47+ providers with confidence scoring

A3M Router is an open-source LLM router that runs queries across 47+ providers in parallel, scores responses by confidence, and returns the best result.

Why A3M?

Parallel Execution — Fires queries to multiple providers simultaneously and scores each response using confidence-based ranking.

EXP3-Inspired Diversity — Uses adversarial bandit techniques to balance exploration vs exploitation.

Semantic Caching — Stores query embeddings to avoid repeated API calls for semantically similar questions.

47+ Providers — OpenAI, Anthropic, Groq, Gemini, OpenRouter, Mistral, and 40+ more.

Local LLM Support — Route to Ollama, LM Studio, or local models for private inference.

Quick Start

npm install adaptive-memory-multi-model-router
npx a3m-router "your question here"

GitHub

github.com/Das-rebel/a3m-router